4 



An index rating of 1.0 Indicates that the food plants so 

 rated were used approximately in proportion to their abundance. An 

 index rating greater 'than 1.0 indicates that the food was preferred 

 and .sought by ducks; of less than 1.0, that the food was less popu- 

 lar than- it was- abundant. 



. . ■ . POSSIBLE SOURCES OF ERKOR 



The index rating figure may or may not 'be a precise indi- 

 cator of the relative value of a waterfowl food plant. Habits of 

 '.".•atprfowl, plant differences and habitat characteristics are factors 

 .tending to prevent a minutely exact rating figure. 



• ■ Ducks frequently feed in, more' tlian one lake in the same 

 day.' The- ar'ea in which a duck is. killed may not be that in which it 

 has filled its gizzard, and yet the mechanics of this study have 

 made necessary the tacit assuitiptipn that tlq contents of each giz- 

 zard were -gathered in - the area in which the duck was killed. 



' •' ■ .Accurate mcasurem.ent of the difrerencos in food yields of 

 the saftie species o.f plant under varying conditions has not been 

 .possible in this- study. For example, sago pondv/eed is very erratic 

 in seed production. Some beds produce no seed; ot}iers produce a 

 large ^amount. In the case of several spt;cies, seeds were available 

 to ducks in an area in which no plants v.erc seen; the seeds liad been 

 dropped in the -mud by t^iC previous year's plants but had not 

 sprouted to produce vegetation. Then, too, no accurate measurement 

 has 'been possible of the effect of differences in water level on the 

 availability to ducks of plant parts. Lovi/ v/atur, or no water, may 

 preclude the use of certain food plants; y/ater of greater than usual 

 depth may have the same efiect \¥ith rcspoct to other plants. 



The index rating in each case should be considered rela- 

 tive rather than exact for the reason that no numerical values were 

 recorded for abundance of certain plants, and these plants were dis- 

 regarded in calculating tl,L.e per cent of _ abundance, of other plants, 

 W'iereas all plants wt,re Included in calculating the per cent of use. 



Even though the possibility of error in Individual ducks, 

 or Ln individual areas, is great, the authors believe that in those 

 food plants in wl.lch the number of samples is large the errors tend 

 to cancel each, other. Field observations confirm or only slightly 

 modify the f indln' s presented .in table 1. Th^e index rating column 

 seems sufficiently valid to justify its use in planting programs, at 

 least for the- eight pl.ants with the highest percentage of use. Of 

 the first eight, those with a low index rating should be avoided in 

 m.ost planting programs. 



AREAS Ul'JBER INVESTIGATION 



Most of th'e duck gizzards collected in 1938 were from but 

 4 areas of the 20 that' were mapped. T.herefore, the per cent of 

 'abundance cf aquatic plants- on only thuse four areas has been con- 

 sidered in this study. • These sample areas contained representatives 

 of all important plants ' existing in the Illinois Ri'vur Valley, under 

 three different types of v;ater levels. 



